Many modern computer systems, particularly enterprise-class systems, are provided with redundant components to provide continuous uptime in the event that one of the system's components fails. In many such systems, the redundancy can be provided by including a pair of identical components which perform the same function in parallel. In the event that one of the redundant components fails, the other component will perform the function alone. This transition from one component to another can be referred to as a “failover.” In other systems, the redundant component is treated as a spare component and is not utilized until after the primary component fails.
Storage systems, for example, may be provided with dual data paths for receiving I/O data from an application host. One such system is the StorEdge T3 Array** sold by Sun Microsystems, Inc., of Palo Alto, Calif. In the event that one of the data paths fails, due to a failure in, for example, the host adapter card on the application host system, the Fibre Channel (“FC”) cable, or FC port on the StorEdge T3 Array, the other data path will handle all of the I/O traffic with the host application.
** StorEdge is a registered trademark of Sun Microsystems, Inc. 
Existing storage systems can monitor redundant data paths and automatically manage the failover to a redundant data path in the event of a failure. However, when a component in these storage systems fails, the system may lose its redundancy protection. Therefore, if the corresponding failover component subsequently fails and there are no other redundant components available, the entire system may fail. Systems which simply manage the failover to the redundant data path but do not adequately warn the system operator of the failure may be left in a non-redundant state for an extended period of time. This increases the chances that the failover component will also fail before the first failed component is replaced. Accordingly, it may be desirable to indicate the first failure to a user in time to enable the user to take steps to replace the failed component promptly before any further failures occur.
In addition, it may also be desirable to warn the user of the failure in an immediately apparent way. Some systems may record the failure in an error log. This error log may be easily accessible to the system operator, but it is unlikely that the operator would review the error log without a more prominent warning. Furthermore, a textual message on a computer screen may not have sufficient visibility to alert an inattentive operator.